tsicl.utils.data_adapter.validate_covar_inputs

tsicl.utils.data_adapter.validate_covar_inputs(covars: list | torch.Tensor | numpy.ndarray | None) → Tuple[bool, list | torch.Tensor | None]

Check dims and standardize all covariate tensor shapes.

Parameters:

covars (list | torch.Tensor | np.ndarray | None) – Covariates of shape (T, C) or (C, T, 1) (possibly with an extra batch dimension)

Returns:

• is_cov_tensor (bool) – True if returned covariate is a batch Tensor.

• covar (torch.Tensor or List[torch.Tensor] or None) – Tensor of covariates of shape (bs, C, T, 1) if all aligned, other a list of covariates of shape (C, T, 1).